We will conduct a nested case-control study of women with benign breast disease to determine the effect of oncogenic and histologic factors on breast cancer risk. The study cohort will consist of all women without a prior history of breast cancer who underwent benign breast biopsies at Vanderbilt Hospital between 1959 and 1988, at St. Thomas Hospital between 1964 and 1988 or at Baptist HOspital between 1982 and 1988. Biopsy slides and paraffin embedded biopsy tissue is available for all of these women. The slides of the entry biopsies will be reviewed without knowledge of subsequent cancer outcome. They will be classified according to Page's histologic criteria (Dupont and Page, N Engl J Med 1985; 312: 146-51). The study cohort contains approximately 10,913 women. We expect to obtain follow-up on at least 85% of these women. Case patients for this study will be cohort members who develop breast cancer during follow-up (approximately 586 women). Two matched controls will be chosen for each case. Control subjects must be cohort members who were at risk of developing breast cancer when their matched case patient developed this disease. Case and control patients will be matched for age, year, benign histologic diagnosis and hospital of their entry biopsy. We will use in situ hybridization techniques to look for altered expression, amplification or transcription of specific oncogenes within different types of breast lesions in the entry biopsies of cases and controls. The oncogenes which will be evaluated as evidence accrues suggesting that they may have a role in the pathogenesis of breast cancer. Sufficient material is available for over 50 such probes on each biopsy. Conditional logistic regression analysis will be used to assess the individual and combined effects of oncogenic, histologic and epidemiologic risk factors in this study. Of particular interest is whether the presence of specific molecular markers affect the breast cancer risk associated with various constellations of histologic and epidemiologic risk factors. We will also assess whether changes in medical practice during the 1970's regarding the use of mammography and the willingness of surgeons to perform breast biopsies has affected the prognostic significance of specific benign lesions. This assessment will be performed using hazard regression analyses. This study will permit the combination of modern methods in molecular biology, pathology and epidemiology to assess potentially powerful new markers of breast cancer risk. This research may lead to important advances in identifying women who can benefit from prophylactic therapy, and may provide information that will prove useful in unraveling the etiology of this disease.